Monday, April 26, 2010

Single Subject Time Series Designs

The single subject time series design (SSTSD) is a form of a case report that uses a more formal process to track the response of a single patient over time. One of the challenges physicians face is the knowledge that it is never really possible to know, for a given single patient, that what we have done to the patient has resulted in the response we see. There can be many reasons a patient responds after therapy, such as natural history, the intervention, placebo effect, and so on. However, the SSTSD provides us with a better window into the usefulness of our intervention, since it tracks a single person over time.

In an SSTSD, you would initially establish a baseline on your patient over a period of at least 3 visits. Here, you would collect data on the patient’s status. For example, perhaps we were working with a patient who has chronic low back pain, and our outcomes measures include an Oswestry Disability Index as well as a Visual Analogue Scale. Taking these three measures gives us a stable sense of the patient’s status. Then, we begin treatment, and during the period of treatment we continue to take these same outcome measurements. We do so for the same length of time as we did for the baseline period. We might see a change in scores when we do this, a reduction in both disability and pain. This can help establish a relation between our intervention and the patient response. Now, let’s add one more period of measurement, a follow-up period where we cease therapy again, and instead measure the outcomes over a similar period of time. If we see the scores rise or return to where they were before we started treatment, this is even better data demonstrating the impact that our therapy has. This is known as an ABA design for a SSTSD.

Conditions ripe for study as SSTSDs need to be stable and chronic; acute conditions vary too quickly for a physician or researcher to be able to establish a meaningful baseline for comparison, because as you can see, the patient acts as his or her own control during the treatment and follow-up time. These kinds of studies are useful for helping a patient to understand change over time, especially for conditions where change in made only incrementally and slowly (for example, frozen should or adhesive capsulitis).

Variations in this design exist. For example, an ABA design can be repeated over time as an ABAB design, also known as a withdrawal design since it withdraws and then restores therapy several times. There can be an ABAC design, where C represents a second, alternative intervention. We can also consider a multiple baseline design, in which at least 3 individuals are tracked like this, but with the caveat that we do not begin treatment on the second patient until after we see a response in the first, and we do not begin treatment on the third patient until we see a response in the second. The differing baselines here are a means to reduce effects of confounding influences. There is also a simultaneous baseline design in which we do the same as in a multiple baseline approach but gather all our patients at the same time; in a multiple baseline approach it is not necessary to do so.

This methodology can be useful in a practice setting, and can be a means for a chiropractor to collect data for future publication. It also is a means to strengthen the results drawn from case reports, which tend to fall lower on an evidence hierarchy. I do not neglect case reports, though; they are often the papers clinicians turn to first, and they can offer direction for challenging patients.

Monday, April 19, 2010

Theoretical Frameworks for Academic Dishonesty

An interesting chapter by DiPietro in the new book “To Improve the Academy, Volume 28” (1) discusses a variety of theoretical frameworks used to explain academic dishonesty. He lays out five such theories, and then places student behavior in their contexts. These theories include deterrence theory, rational choice theory, neutralization theory, planned behavior theory and situational ethics.

1. Deterrence theory: this theory proposes that cheating is a function of the severity of the consequences. Thus, if we want to prevent or stop certain behaviors, we need to punish them with consequences so severe it will act as a discouragement. Such punishments might include failing the assignment or course, probation or expulsion. This is based on past research demonstrating that when people believe they can engage in a behavior with no or minimal consequences, they are likely to do so. One of the chief challenges here is that, due to the increased time and effort involved, instructors may not wish to report the behavior. There are also cultural determinants; Western students fear expulsion, while Asian students seem to fear public humiliation.

2. Rational Choice theory: Here, this theory treats dishonest actions as the result of decisions that we make as rational agents; that is, we weight pros and cons of an action, and based on how we assess the alternatives, we make our choice. We might look at this as a kind of cost-benefit analysis: is the effort necessary to cheat worth the cost of getting caught and being punished?

3. Neutralization theory: This theory hypothesizes that students are able to engage in morally inappropriate acts without damage to their self-esteem if they are able to rationalize the act and consider it morally neutral rather than wrong. I think of a recent news report about a young author charged with plagiarizing from an existing text who stated that she was using the material as a “mash-up”; that is, taking someone else’s words and placing them into he rown work as part of her creative process. If we can convince ourselves that what we are doing is not morally wrong, we may then proceed with doing the actual act. Thus, efforts to prevent cheating should work on deneutralizing it, emphasizing the moral incorrectness of the act. We should focus on personal responsibility.

4. Planned Behavior theory: This theorizes that cheating happens because of the opportunity as well as the intention to treat. Thus, we need to take efforts to reduce the opportunity to cheat, perhaps be increasing our vigilance during exams, using additional proctors and exam versions, and by increasing education on the value of integrity and honesty. Example: place open seats between students; this reduces opportunity.

5. Situational Ethics: This appears to be related to rational choice theory and is a direct outgrowth of John Stuart Mills and his initial utilitarianism. That is, each student has to weight the specifics of his or her situation; could I cheat here because of these reasons? Can I accept the risk knowing the potential benefit? What are the issues I am concerned with here? If I do poorly, I could lose my scholarship and be sent home; thus, maybe I need to do something to increase my chance of passing, and this might include cheating. It is hard to address this because each student will bring his or her specific issues to the consideration.

These approaches help us understand something about why students cheat, but we do need to ensure consistency in our approach and in our investigations of possible cheating. None of us are naïve about this; we know it happens and we take what measures we can. Moral training and a focus on the integrity of being a doctor may help us to diminish the cheating that undoubtedly occurs.

1. DiPietro M. Theoretical frameworks for academic dishonesty. In: Nilson LB, Miller JE. To Improve the Academy, Vol. 28. San Francisco, CA: Jossey-Bass, 2010:250-262

Monday, April 12, 2010

Innovation in Education- Amazing Examples

I would like to begin this entry in my blog by asking you to watch the following video:

When you do, focus on the following. Note how the children involved are engaged in the process of what they are doing, how safe they have to feel in order to express themselves in the way that they do, what it must have taken to get a group of children in a public school to work that hard and to work together, and the commitment of the young man playing the piano, who leads these children in their singing. That man, Gregg Breinberg, came to music education and decided that maybe there was a new way to engage children, one that did not rely upon the so-called “school band” sound that pervades music training for young children. The standard pedagogy focuses upon old standards, such “America the Beautiful” and “Michael, Row Your Boat Ashore.” And while these are indeed fine songs, they may no longer really speak to children as well as they might. So Breinberg tried something new; he decided to develop vocal harmony in modern rock and soul songs. This meant teaching his kids to sing songs by Talking Heads, Rihanna, Lady Gaga, Stevie Nicks, Tori Amos, and in the case of this video, the rock band Phoenix.

Imagine being a kid in Staten Island, where these kids are located. Imagine learning to sing a song by someone like Tori Amos, likely a singer you might never have heard of, but who is a modern pop singer. So you learn the song. And then you are asked to travel across the river to sing a short set of songs in New York City, and lo and behold, when you get there, so does Tori Amos, who then sings with you!

This is innovation in action. Mr. Breinberg had formal training as a music educator, knew the standard methods for teaching music, and out of that created a new way of teaching children to sing in choir that stood the world on its head. Yet his success is undeniable and reading his blog, linked immediately above, can make you feel much better about education in general.

Then, look at this little clip of MD Weathers teaching math at Biola College:

Think about how long it must have taken him to set this up. How committed he has to be to his students to invest the time and the creativity to his teaching efforts. Yet again, this is an example of a teacher moving beyond standard pedagogy, trying something new and novel, something humorous and yet engaging. In both of these teachers we see efforts to keep students engaged, to keep them just a little bit off balance, to ensure they participate with what is going on in the classroom. And the question for all of us, what can we do that might shake up a standard classroom setting? Can we bring in media? I am experimenting with using the TV program House to teach examples of bioethics (or in House’s case, usually, violations of bioethics). Can we try new methods? What might those methods be? Think about it, and give it a shot; it will be enjoyable for you and exciting for your students.

Tuesday, April 6, 2010

Using Clinical Trial Results

Even the best randomized clinical trials cannot tell us for whom a given treatment is better, only that a given treatment is better than what it is being compared to. When we then read that trial after it has been published, we have to make a determination about whether or not we can apply it to our patient or to our specific set of circumstances. How can we go about doing so? Alejadro Jadad, developer of an instrument used to assess the quality of randomized trials, offers some thoughts on this (1). He suggests we consider the following:

1. Does the research question match your own question? This is critically important. All research studies should provide you the question, which should also speak about the characteristics of the participants, the condition, the setting, the interventions, and the outcomes. It is surprising how many papers never actually present the actual question, but force you to read over the methods to determine what exactly is being studied.

2. Does the report include enough information on the execution of the study? This will help tell you how well it was executed and whether or not you are comfortable that the researchers could answer the question they have asked. Issues to consider include:

(a) What was the sampling frame? How were prospective patients approached by investigators? We wish to know something about the population; there is a difference, for example, between patients seen in emergency room settings and those in private practice settings. The question about how patients were approached gets at the population sample and it generalizability: was this a convenience sample, a random sample, or something else.

(b) What were the inclusion and exclusion criteria? Again, this is to address issues of external generalizability. This gives information on the health status of participants, potential co-morbidity, etc.

(c) Was the setting appropriate? Was it similar to your own setting? See above; some settings have patients bases significantly different from the ones you see in practice, and as a result you may not feel comfortable generalizing the information from the study results to your own patients.

(d) What were the interventions and who gave them? We see many trials involving spinal manipulation, but certainly not all are provided by chiropractors. Some might have been given by physical therapists, whose training, philosophy and approach differs from ours. Thus, perhaps we are less comfortable in generalizing results.

(e) Was there randomization and blinding? This helps ensure less bias in results and better comparison between groups.

(f) What were the outcomes of interest? This gets at whether or not you can use the results to inform your own patient management question. These should matter both to you and to the patient ( a so-called “patient-oriented evidence that matters” or POEM).

This approach can help you determine whether or not the paper you are reading can be used to help decide what to do with the patient you are caring for. Jadad nicely illustrates these points in his informative little book.

1. Jadad A. Randomised controlled trials. London; BMJ Books, 1998:69-73